摘要
为了解决缓变故障的检测和容错问题,针对容错设计中小幅度增长的缓变故障,考虑组合导航系统导航传感器的特点,提出了一种新的组合导航缓变故障容错方法。定义了滤波器观测信息的故障因子,提供了基于自回归条件异方差模型的故障因子模糊评估方法,并利用故障因子来自适应调整观测噪声协方差阵中相应的量值,从而达到抑制缓变故障的作用。通过仿真分析和实验验证了缓变故障容错方法的有效性。实验结果表明,在导航传感器发生缓变故障时,这种缓变故障容错方法可以有效地改善组合导航滤波精度。
Addressing the issue of fault detection and fault-tolerance for gradual changing faults, a novel fault-tolerance method of integrated navigation system was proposed for gradual changing fault with smaller extent of difficult fault-tolerance by considering the unique characteristics of the navigation sensor. The concept of fault factor and its fuzzy evaluation based on autoregressive conditional heteroskedastic（ ARCH） model were presented, and then fault factor used to correct the observation noise covariance to prevent gradual changing fault from affecting the filter. Simulation and experimental analyses show the effectiveness of the method which can effectively improve the estimating precision of the integrated navigation system when gradual changing faults exist in system.
出处
《解放军理工大学学报（自然科学版）》
EI
北大核心
2014年第6期519-526,共8页
Journal of PLA University of Science and Technology(Natural Science Edition)
基金
国家973计划资助项目(2011CB707004)
国家自然科学基金资助项目(61101077)
国家科技支撑计划资助项目(2011BAH24B02)
关键词
ARCH模型
组合导航
容错
缓变故障
ARCH model
integrated navigation
fault-tolerance
gradual changing fault
